INFORMAL LOGIC
XVI.l, Winter 1994
Book Review
Testimony: A Philosophical Study by C. A. J. Coady
CAROL CARAWAY Induma University of Pennsylvania
Coady, C. A. 1. (1992). Testimony: A
Philosophical Study. Oxford: Clarendon
Press. pp. x, 1-315. ISBN 0-19-824786-9.
Cloth. US $39.95.
With a few noticeable exceptions, phi-
losophers in epistemology, logic, and phi-
losophy of language have generally
ignored testimony. Empiricist epistemolo-
gists have focussed on the individual
knower's sensory perceptions and ignored
information gained from what others say.
Logicians generally discuss testimony
briefly in connection with ad hominem ar-
guments, pointing out that attacking a per-
son's character or credentials is not a
fallacy if we are evaluating testimony rath-
er than argument. For while argument must
be evaluated on its own merits (validity
and soundness), testimony must be evalu-
ated based on the competence and honesty
of the witness.
In Testimony: A Philosophical Study,
C. A. J. Coady seeks to put an end to phi-
losophers' unjustified neglect of testimony.
He argues that trusting another's word is
"fundamental to the very idea of serious
cognitive activity" (p. i). Coady divides his
presentation into five sections: L The
Problematic, II. The Tradition, III. The
Solution, IV. The Puzzles, and V. The
Applications.
Section I. The Problematic explores
the extent, depth, and inevitability of our
reliance on what others tell us. Most epis-
temologists have either ignored testimony
altogether or rejected it summarily. Coady
argues that this tradition of neglect is un-
justified and harmful because our reliance
on testimony is crucially important to both
our ordinary beliefs and to theoretical pur-
suits, such as history, psychology, and
physics. Psychology and medicine, for ex-
ample, rely heavily on subjects' verbal re-
ports. Moreover, experts and researchers in
all the sciences use the testimony of other
researchers.
Coady offers two primary reasons for
epistemologists' neglect of testimony: the
individualist ideology of such thinkers as
Descartes and Locke, and the association
of testimony with Medieval reliance on the
opinions of authorities.
Coady sketches four theoretical re-
sponses to realizing the extent of our
reliance on testimony.
1. the Puritan Response: Testimonial
beliefs are not known; therefore,
knowledge is much rarer than we
think (Plato & Collingwood).
2. the Reductive Response: Testimo-
nial beliefs are known and infer-
entially justified (Hume, Russell,
W. K. Clifford & J. L. Mackie).
3. the Fundamentalist Response:
Testimonial beliefs are known and
non-inferentially justified or basic
(Thomas Reid).
4. the End-of Epistemology Response:
Testimonial beliefs are not infer-
ential, and positive epistemology,
62 Carol Caraway
i.e., foundationalism, is a radical
mistake (Quine).
Coady's list embodies the questionable
assumption that foundationalism is the
only plausible theory of the structure of
justification and knowledge. Coady sees
Quine as signaling the end of epistemology
by proclaiming the failure of foundational-
ism. Coady wants to preserve epistemolo-
gy by defending foundationalism. But if
foundationalism is not the only option,
then its alleged death need not mean the
end of positive epistemology. Indeed, con-
temporary American epistemologists have
offered strong defenses of two other theo-
ries: coherentism (Keith Lehrer, Larry
Bonjour, and Lynn Hankinson Nelson)
and contextualism (David Annis and Helen
Longino). Later, I shall argue that Coady's
ignoring these alternatives to foundational-
ism leads to serious confusion. He com-
pares himself to Reid, yet gives arguments
that seem more appropriate for a coheren-
tist (or perhaps even a contextualist). Fur-
ther confusion arises because traditional
foundationalism assumes an individualist
approach which Coady rejects.
Coady defines two concepts of testi-
mony: formal or legal testimony and natu-
ral or everyday testimony.
Formal Testimony [FT] has the follow-
ing marks:
(a) It is a form of evidence.
(b) It is constituted by persons A of-
fering their remarks as evidence
so that we are invited to accept p
because A says that p.
(c) The person offering the remarks
is in a position to do so, i.e. he
has the relevant authority, com-
petence, or credentials.
(d) The testifier has been given a
certain status in the inquiry by
being formally acknowledged as
a witness and by giving his evi-
dence with due ceremony.
(e) As a specification of (c) within
English law and proceedings in-
fluenced by it, the testimony is
normally required to be firsthand
(Le. not hearsay).
(f) As a corollary of (a) the testifl-
er's remarks should be relevant
to a disputed or unresolved ques-
tion and should be directed to
those who are in need of evi-
dence on the matter (pp. 32-33).
I find FT condition (f) misleading. Giv-
en the possibilities of epistemic overdeter-
mination and counterevidence, there may
be times during the court proceedings
when the jury already has enough evidence
to reach a verdict and so is not really "in
need of evidence on the matter." I would
modify (f) to state: the testifier's remarks
.. , should be directed to those who must
evaluate the evidence on the matter to
reach a verdict. I am also confused about
the status of FT (a) (f). Coady refers to
them as "marks of formal testimony." This
characterization is vague. Are they individ-
ually necessary and jointly sufficient con-
ditions for FT, or are they something else,
such as Wittgensteinian criteria? Mark (e)
concerns what is "normally required" and
so seems to be something less than a nec-
essary condition.
Two additional marks which Coady ex-
cludes from his definition of FT are (g)
creditability or sincerity and (h) corrobora-
tion. He excludes (g) creditability because
insincere testimony is still testimony and
because a sincerity condition is not unique
to testimony. He rejects (h) corroboration
because although ancient Roman Law re-
quired it, British Law does not.
Having defined Formal Testimony
[FT], Coady turns to his primary concern:
Natural Testimony [NT]. Here, he finds
parallels with FT conditions (a), (b), (c)
and (f), but not with (d) and (e). He defines
Natural Testimony [NT1 as follows: A
speaker S testifies by making some state-
ment p if and only if:

( I) His stating that p is evidence that p
and is offered as evidence that p.
(2) S has the relevant competence,
authority, or credentials to state
truly that p.
(3) S's statement that p is relevant to
some disputed or unresolved
question (which may, or may not
be, p?) and is directed to those
who are in need of evidence on
the matter (p. 42).
The two competency conditions Ff (c)
and NT (2) differ noticeably. Ff (c) re-
quires that the person offering the remarks
be in a position to do so. NT (2) seems
stronger, for it requires S to be in a position
"to state truly that p." This apparent
change in the competency condition is
confusing. NT conditions should be weak-
er than Ff conditions, not stronger. Re-
quiring S to be competent to state truly that
p seems too strong. Coady wants to allow
that so long as S's remarks conform to
probability and explanatory conditions
such as those spelled out by Achinstein
(The Nature of Explanation, Oxford,
1983), S has provided evidence for p even
if p turns out to be false (pp. 44-45). I sug-
gest replacing NT (2) with what I will call
NT (2)*: S has the relevant competence,
authority, or credentials to state justifiedly
that p (or to provide evidence that p).
Another problem with competency
conditions Ff (c) and NT (2) is the implicit
inclusion of sincerity. Coady excluded a
sincerity condition from Ff for reasons
that also apply to NT. Yet, in discussing
NT (2) he explicitly discusses trust, hones-
ty, and deceit. This confusion is clarified
by his earlier remark, "I argued against the
inclusion of a credit condition in our defi-
nition of formal testimony but the cases of
the insane and the very immature [young
children] suggest that a capacity for sin-
cerity might very well be part of the cre-
dentials we require under condition (c)" (p.
36). Now, while I agree that the capacity
Review of C. A. J. Coady 63
for sincerity may be included in compe-
tence, I believe Coady should supplement
his definitions of Ff and NT with defini-
tions offelicitous testimony which include
a sincerity condition. When we evaluate a
witness, we must evaluate both her compe-
tence and her sincerity. Coady's initial
analysis should prepare the way for such
evaluation.
Indeed, I would prefer a somewhat
more radical revision: constructing three
definitions of "natural testimony" that par-
allel the definitions of "argument;' "valid
argument," and "sound argument." "Natu-
ral testimony" would include only NT (1)
and (3) and WOUld, thus, require neither
competency nor sincerity. "Valid" natural
testimony would require NT (2)* compe-
tence, but not sincerity; and "sound" natu-
ral testimony would require both NT (2)*
competence and sincerity. While I support
Coady's using legal testimony and Austini-
an philosophy of language to illuminate
certain aspects of natural testimony, I be-
lieve my "critical thinking" approach
might prove more useful. It would elimi-
nate some of the confusions Coady en-
counters and could also incorporate
insights from the law and from Austin.
In Section II. The Tradition, Coady
explores past philosophical writings on
testimony by Hume. Price. Russell, and
Reid. His analysis reveals the individualist
emphasis which gave rise to the idea of
"autonomous knowledge." Hume, Price,
and Russell all begin with the problem of
how an individual with perception, memo-
ry, and reasoning can determine whether
reports are sufficiently reliable to count as
knowledge and, thus, extend her very
restricted knowledge base.
Chapter 4. Testimony, Observation.
and the Reductive Approach, examines
two versions of the Reductionist Thesis
[RT]. The first is David Hume's view
[RT-H] that testimony reduces to inductive
inference that reduces to observation.
Hume maintains that we believe testimony
because we humans as a group are accus-
64 Carol Caraway

tomed to finding conformity between testi- be adequately supported by individual

mony and reality (An Enquiry Concerning observation alone (pp. 102-104). (A) is
Human Understanding, Oxford, 1957, p. more like a maxim or methodological rule:
113). Coady argues that RT-H is viciously either (B) believe what you are told by oth-
circular and must, therefore, be rejected. In ers unless or until you have reasons for
contrast to Hume, J. L. Mackie bases his doubting it or (C) conduct your thoughts
view [RT-M] not on the observations of and your actions as if A were true (p. 104).
humans as a group, but on the observations Price defends B/C as epistemically
of the "autonomous knower" ["The Possi- expedient in that often we must either rely
bility of Innate Knowledge," Proc. of the on testimony or suspend judgement.
Aristotelian Society (1970), 254]. Mack- Coady compares Price's argument to
ie's autonomous knower is someone who Pascal's wager and constructs the follow-
relies on testimony only when he has ing Testimony Wager.
checked the witness's credibility. Coady
SI S2
rejects RT-M because it is false that indi- Consequence A True A False
viduals do this amount of credibility Matrix:
checking (p. 82).
Al Choose to Greatly Great extension
Coady rejects both versions of reduc- believe extended of error and
tionism. For him, testimonial beliefs are testimony 'true belief' some new true
basic: they cannot be reduced to inference reliable and some new beliefs
and observation. Some counter this with (Adopt C) false beliefs
the fact that observation sometimes leads A2 Choose Great failure to Great safety
us to reject testimony. Coady rightly ar- not to believe get available from error
gues that this fact does not prove that testi- testimony 'true belief' (p. III)
mony reduces to observation. In opposition reliable
to it, he presents two other facts: (Reject C)
Desirability 10 -10
1. One testimony can lead me to re-
Matrix: -10 10
ject another without appealing to
personal observation. Probability .5 .5
Matrix: .5 .5
2. Testimony sometimes leads us to
Expected Desirability of A I = 5 - 5 = 0
reject an observation, e.g. as hal-
lucinatory (p. 97). Expected Desirability of A2 =-5 + 5 0 (p. 113)

Thus, Coady here argues that it is still an
open question whether testimony is reduci-
ble to observation.
In Chapter 5. Deciding on Testimony,
Coady presents H. H. Price's view from
lecture 5 of Belief (London, 1969). Price
argues· that testimony is an important
source of knowledge about a wide range of
things: geography, history, one's own age
and birthday, etc. Our underlying assump-
tion about testimony in the great majority
of cases is "(A) What there is said to be (or
to have been) there is (or was) more often
than not" (p. 102). Although (A) looks like
an inductive generalization, it cannot
Coady concludes that when we compare
the risk of being in error with the possibili-
ty of gaining more true beliefs, there is no
reason to choose one over the other. We
rely on testimony and will continue to do
so, but our reliance is unjustified by the
arguments used or implied by Price.
Chapter 6. The Analogical Approach
examines Bertrand Russell's theory of tes-
timony. For Russell, I know that other
minds exist and are communicating inten-
tions by analogy with my own case. But
can analogical argument from my own
case guarantee the general sincerity and
veracity of testimony? Russell does not

explicitly use analogy in this way. F. H.
Bradley does. He argues that "our belief in
testimony is justified by an inference to a
mental state in the witness essentially one
with our own" (p. 118). I can trust anoth-
er's testimony only if I can view her as ob-
serving on my behalf, and I can so view
her only if she has "the same mental out-
look" I do (p. 118). Coady argues that
since identity of outlook cannot be deter-
mined without relying on testimony,
Bradley's approach is circular and will not
solve Russell's problem. Thus, on Coady's
analysis, the theories of testimony devel-
oped by Hume, Price, and Russell all fail.
Coady finds more promising Thomas
Reid's view that testimonial beliefs are not
inferential, but basic.
In Chapter 7. Scottish Fundamental-
ism, Coady explores Reid's analogy be-
tween testimony and perception. Reid
rejects the standard view that testimonial
beliefs are inferential, maintaining that
normally we accept what we are told as
reliable just as we accept "the testimony of
our senses" (p. 123). Both perception and
testimony require the proper functioning
of the relevant cognitive mechanism(s) and
the satisfaction of certain conditions.
Neither requires that S know what the
mechanisms are or that the conditions
are satisfied. Thus, for Reid, testimonial
beliefs are basic, as are perceptual beliefs.
In Section III. The Solution, Coady
argues that testimony is a basic source of
knowledge. He begins by examining simi-
larities and differences between testimony
and other sources of information. Unlike
perception and memory, testimony is al-
ways testimony that. This, he argues, en-
tails neither that perception is more basic
than testimony nor that there is no basic
testimonial knowledge. Some have argued
that testimony is inferential because all our
testimonial beliefs are mediated by the
witness's veracity, reliability, etc. Coady
rejects this argument as fallacious because
it does not fit our actual practices. We do
Review of C. A. J. Coady 65
not ordinarily consider, and often cannot
investigate, the witness's veracity, reliability,
etc. Moreover, the same reasoning applies
to perception. Perceptual beliefs rest on
truths about the conditions of perception,
but ordinarily we need not establish these
truths as part of the justification of our per-
ceptual beliefs (pp. 141-144).
Nonetheless, perception may still seem
more basic than testimony, for testimony is
possible only via perception: S must hear
the testimony or see it written. Coady in-
sists that this does not make testimony in-
ferior to perception:
The giving and receiving of testimony is, as
Reid calls it, 'a social operation of the
mind'. which presumes upon perception in
the ways already indicated but which has
its own epistemic autonomy (p. 147).
Coady prefers to think of perception,
memory, testimony and inference as all on
the same level, but with perception in the
center. These four sources of information
interpenetrate each other. Memory, testi-
mony, and inference enter essentially into
perceptual belief formation. There is no
perception so pure that it is uncontaminat-
ed by memory, testimony, and inference.
Consequently, the idea of making percep-
tion the basis of memory, testimony, and
inference is absurd (pp. 146-147).
I see a possible confusion in Coady's
reasoning here. Granting perception and
testimony equal status entails that testimony
is basic only on the foundationalist as-
sumption that perception is basic. Coher-
entism gives perception and testimony
equal justificational status: both are infer-
ential or non-basic. In making inference an
essential part of perception, Coady seems
to be committed to the coherentist view
that perception is inferential, not basic.
The "hankering after a primacy for per-
ception is really a hankering after a prima-
cy for my perceptions" (p. 148). This,
Coady argues, assumes the egocentric
starting-point of traditional epistemology:
"the epistemically isolated self' (p. 149).
66 Carol Caraway
Coady maintains that this starting-point is
"a product of cultural and philosophical
predilection rather than an a priori inevita-
bility" (p. 149). Within it he sees the seeds
of another. For Descartes's view of the in-
dividual's native powers leads into "the
idea of the subject as endowed with pow-
ers and capacities as a member of a species
so endowed" (p. 150). This, Coady argues,
should shift the focus from the individual
to the community. Thus, "we see our start-
ing point as encompassing our knowledge
and not exclusively my knowledge" (p.
150). We should begin epistemology by
asking, "What is the nature of our knowl-
edge?" or "How do we come to know?" (p.
150) Unfortunately, Coady does not dis-
cuss what forms communal theories of
knowledge can take. Can there be a com-
munal foundationalism, or must commu-
nalism be either contextualist or
coherentist? Is Coady's theory a commu-
nal foundationalism? If so, he fails to make
explicit the nature of such a view.
Coady insists there is a "fundamental
flaw" in individualist justifications, such as
Mackie's, for communal epistemological
trust. Such justifications covertly depend
on the reliability of trusting others while
overtly trying to establish that reliability
(p. 152). This dependence is inevitable be-
cause such justifications must begin by as-
suming a public language, and to
understand such a language, I must treat
many of the reports expressed in it as true.
Can any argument provide a justifica-
tion or philosophical rationale for our very
extensive trust in testimony? In answer to
this question, Coady offers an intellectual
overview of our testimonial practices. He
begins by exploring Donald Davidson's
project of radical interpretation ["Radical
Interpretation," inquiries into Truth and
interpretation (Oxford, 1984), pp. 125-139].
Coady believes that Davidson's fundamen-
tal insight is one shared by the later Witt-
genstein: understanding another's speech
requires a shared w6rld view (Davidson,
"The Method of Truth," inquiries, p. 199).
Coady states Davidson's first principle
of charity thus:
(a) We must apply a principle of charity (or
some similar interpretive maxim) in inter-
preting the speech of others, most notably
an alien community, so that agreement is
maximized or optimized amongst us and
them. We must, that is, find their expressed
beliefs mostly correct by our lights (pp.
157-158).
Coady examines Davidson's three argu-
ments for (a) and Colin McGinn's criti-
cisms of them ["Charity, Interpretation and
Belief," The Journal of Philosophy, 74
(1977),521-535]. The third argument con-
tains the best reconstruction of what ap-
pears most promising in the first two
arguments (p. 162). It may be paraphrased
as follows:
1. We cannot identify beliefs at all
without locating them in a wider
pattern of beliefs most of which
must be true.
2. The pattern determines the
subject-matter of the belief.
3. If we cannot identify beliefs, we
cannot identify meanings, since
beliefs/desires/meanings are de-
livered together in the process of
interpretation.
Therefore, interpreting speech re-
quires the general correctness of
belief (p. 162).
McGinn offers the following coun-
terexample. Some ancient peoples believed
that "the stars were apertures in a vast
dome through which penetrated light from
a great fire beyond" (p. 162). We can iden-
tify such a belief and related beliefs with-
out holding them true. Indeed, we
recognize that such beliefs are radically
false. What allows us to identify such a be-
lief as a belief about the stars is not that it
is true, but that we can use our concepts
and our (true) beliefs to identify the
objects it was about (pp. 162-163).

Coady concludes that McGinn's criti-
cisms of Davidson's argument show not
that we can dispense with a principle of
charity, but that we can tell what some be-
liefs of an alien community are about, and
even what many such beliefs are, without
treating those beliefs as shared or true.
Thus, Coady concludes, (a) is too strong,
but contains a solid core of truth. We must
assume that aliens inhabit the same physi-
cal universe, evolved in broadly similar
ways, and need nourishment, reproduction,
safety, cooperation, etc. This commonality
of constitution produces some basic simi-
larity of outlook and therefore, a consider-
able commonality of beliefs and interests
(pp. 166-168).
Coady concludes Section III. by reiter-
ating that we do not prove from purely in-
dividual resources that testimony is
reliable. Rather, beginning with an inevita-
ble commitment to some degree of reliabil-
ity, we find this commitment strongly
enforced and supported by the cohesion
between our informational sources and the
coherence of our beliefs (p. 173). Coady
characterizes his position as fitting none of
the four outlined in Chapter 1. He sees his
position as most like the fundamentalist
approach of Reid, but without relying
merely on an appeal to intuitively evident
first principles. Mistakenly assuming that
foundationalism is the only plausible theo-
ry, Coady makes testimonial beliefs basic.
Had he seriously considered the alternative
theories of coherentism and contextualism,
his conclusion might have been very dif-
ferent. Coherentism puts testimony and
perception on the same level by making
both inferential. Coady's remarks about
cohesion and coherence together with his
earlier remarks about perception and infer-
ence seem to point to some type of coher-
entist or to a mixed foundationalistl
coherentist theory. Yet, he seems unaware
of this implication.
In Sections rv. and V. Coady extends
his analysis to other issues. Section IV.
The Puzzles deals with such paradoxes as
Review of C. A. J. Coady 67
Locke's view that testimonial transmission
must lead to less and less reliability of the
transmitted message, and hence to the
eventual disappearance of history. In Section
V. The Applications, Coady shifts towards
more applied issues and uses his new view
of testimony to challenge assumptions in
history, mathematics, psychology, and law.
In Chapter 13, for example, Coady gives
counterexamples to R. G. Collingwood's
influential denial of any role for testimony
in serious historical research. In Chapter 15,
he criticizes claims of psychologists that
testimony is unreliable. Psychological
researchers often rely implicitly on the
reliability of testimony to prove its un-
reliability. They also operate from three
questionable background assumptions
about the common man and the law:
super-realism, passive recording, and very
great accuracy.
Influenced by Austin, Ryle, Grice,
Anscombe, and Kneale, Coady is steeped
in the Oxford tradition, and Testimony
demonstrates that this tradition is still very
much alive. The book is not aimed exclu-
sively at philosophers. Sections L, II., and
V. can be understood by undergraduate
philosophy majors and well-educated
non-philosophers. Sections III. and IV. are
more technical. Non-philosophers will
have difficulty understanding some of the
sophisticated issues discussed there. Sec-
tion III. is suitable for graduate students in
philosophy, but should be used with under-
graduates only at the end of a course in the
philosophy of language. Testimony could
be used with other texts in several graduate
courses: epistemology, philosophy of lan-
guage, philosophy of law, and philosophy
of history. The use that would do most jus-
tice to the book, however, would be as the
focal text in a graduate seminar on testimo-
ny. It could, then, be supplemented by rele-
vant passages from Plato, Hume, Reid,
Russell, Price, Mackie, Davidson, McGinn,
and other relevant authors.
"Testimony is a prominent and under-
explored epistemological landscape" (p. i)
68 Carol Caraway

which deserves much more attention than scholars interested in testimony, this book
it has received. Coady has given us an im- is a "must read."
portant look at this long neglected topic.
Testimony makes an important theoretical CAROL CARAWAY
PHILOSOPHY AND RELIGIOUS STUDIES
contribution to the field of informal logic DEPARTMENT
by advancing our understanding of judge- INDIANA UNIVERSITY OF PENNSYLVANIA
ments of the credibility of reports. For INDIANA, PA 15705-/087 0